This does most of the work for you, but I've left some things for you to do:

<DATA>;

my @prev_line;
while ( <DATA> ) {
    my @this_line = split /[,\s]+/;
    shift @this_line;
    print join( ", ", map { "$this_line[$_]-$prev_line[$_]" } 0 .. $#t
+his_line ), "\n"
        if @prev_line;
    @prev_line = @this_line;
}
__DATA__
date, val1, val2, val3, val4
1/2/2007, 1, 4, 5, 6
1/3/2007, 2, 5, 7, 10
1/5/2007, 5, 6, 8, 11
[download]

Output:

2-1, 5-4, 7-5, 10-6
5-2, 6-5, 8-7, 11-10
[download]

You may find this version easier to follow (using the same __DATA__ as earlier):

<DATA>;

my @prev_line;
while ( <DATA> ) {
    my @this_line = split /[,\s]+/;
    shift @this_line;

    if ( @prev_line ) {
        my @output;
        for ( 0 .. $#this_line ) {
            push @output, "$this_line[$_]-$prev_line[$_]";
        }
        print join( ", ", @output ), "\n";
    }
    @prev_line = @this_line;
}
[download]

That leaves you with the job and handling the files.

I'm sure google will help, and there's always the Perl documentation.

update: omitted __DATA__ first time round
update^2: added second example